Although our primary hypothesis was that genes in the NPY system were associated with alcohol dependence, we analyzed additional phenotypes related to alcoholism that were suggested by the literature. The withdrawal phenotypes each consist of a subset of the alcohol dependent subjects, as does the phenotype of comorbid alcohol and cocaine dependence; thus they are not independent phenotypes, but phenotypes nested within alcohol dependence, analyzed to better understand what aspect of alcoholism is most affected by these genes. The cocaine dependence phenotype was considered a secondary analysis, and while it includes many subjects who also meet criteria for alcohol dependence, it is not a nested subset. Therefore, we are testing two phenotypes (alcohol dependence and cocaine dependence) and we have considered the strength of our association results if we were to apply a conservative Bonferroni correction (0.05/2 = 0.025). Applying this correction, we would still identify one SNP in NPY2R which is significantly associated with alcohol dependence and with comorbid alcohol and cocaine dependence, two SNPs in NPY5R which are significantly associated with withdrawal with seizures and four SNPs in NPY2R that are significantly associated with cocaine dependence. The results from haplotype analyses of NPY2R are even stronger (Table 3).
